Regex becomes greedy with /s

pingo has asked for the wisdom of the Perl Monks concerning the following question:

Dear monks, I have stumbled across an unexpected regex behaviour that I hope someone can explain. More specifically, this is the code:

$text = "<a href='http://www.example.com/'>Example 1</a>\n<a href='htt
+p://www.example.net/' target='_blank'>Example 2</a>";

my ($a) = $text =~ /<a href='(.+?)' target='_blank'>/s;
print "$a\n";
[download]

And this is the output:

http://www.example.com/'>Example 1</a>
<a href='http://www.example.net/
[download]

I would have expected this, though:

http://www.example.net/
[download]

It is not much of a problem as such, as I can always change .+? to [^']+ or remove the /s. What I find a bit odd is that the /s makes the .+? greedy, despite the +?. Could someone please enlighten me?

Comment on Regex becomes greedy with /s Select or Download Code

Replies are listed 'Best First'.
Re: Regex becomes greedy with /s by Corion (Patriarch) on Feb 26, 2010 at 10:02 UTC
It's not that the `.+?` becomes greedy, it's that `/s` allows dot to match newlines. Thus, the leftmost match can start earlier. Maybe you want to use HTML::LinkExtor or an HTML parser, or maybe even Web::Scraper, or even HTML::Selector::XPath	[reply] [d/l] [select]
Re^2: Regex becomes greedy with /s by pingo (Hermit) on Feb 26, 2010 at 10:24 UTC
Ah... I think I understand. And this is just a throwaway script I am working on (at least I am fairly sure it is ;-)). If it was something more permanent, I'd certainly use one of the modules you mention.	[reply]
Re^3: Regex becomes greedy with /s by AnomalousMonk (Archbishop) on Feb 26, 2010 at 23:35 UTC
Another argument for TheDamian's PBP regex recommendation that a standard set of regex modifiers be used, even in 'throwaway' scripts. E.g., `m{pattern}xms`	[reply] [d/l]